aggregation behaviour in a neotropical dendrobatid frog ... · variety of taxa, amongst amphibians...

Aggregation behaviour in a neotropical dendrobatidfrog (Allobates talamancae) in western Panama

Gareth R. Hopkins1,2,3) & Peter N. Lahanas2)

(1 Department of Biology and the Ecology Center, Utah State University, Logan, UT, USA;2 Institute for Tropical Ecology and Conservation (ITEC), Gainesville, FL, USA)

(Accepted: 3 February 2011)

Summary

Aggregation is a common behaviour in a number of animal taxa and is used for a variety ofpurposes. For example, aggregation may be a response to environmental resources, sexualreproductive behaviour or an antipredator response. Although commonly recognized in avariety of taxa, amongst amphibians aggregation has rarely been reported in adult poison dartfrogs (family Dendrobatidae). The Talamanca striped rocket frog (Allobates talamancae) isa small, non-toxic leaf-litter dendrobatid frog found in the Bocas del Toro archipelago ofwestern Panama. The purpose of our study was to describe the dispersion pattern of Allobatestalamancae and elucidate a possible explanation for this pattern. Based on observationsof grouping behaviour in captive frogs, we hypothesized that A. talamancae significantlyaggregate in natural habitats. We conducted a series of field-based experiments that indicatethat this species does significantly aggregate in the wild. Preliminary results suggest thataggregation may be an adaptive response to predation risk. We discuss this possibility, aswell as critique other explanations for aggregation in this species, placing these results in thebroader theoretical context of aggregation as an adaptive behaviour in animals.

Keywords: aggregation, antipredator, Dendrobatidae, Anura, frog, Allobates, Colostethus,neotropics.

1. Introduction

The aggregation of animals in multi-individual groups is a widespread be-havioural phenomenon observed in a diversity of taxa. Perhaps most well-known are large congregations of migrating mammals (e.g., Fryxell, 1995),

3) Corresponding author’s e-mail address: [email protected]

© Koninklijke Brill NV, Leiden, 2011 Behaviour 148, 359-372DOI:10.1163/000579511X559607 Also available online - www.brill.nl/beh

360 Hopkins & Lahanas

flocks of birds (e.g., Miller, 1922; Heppner, 1974), or colonies of social in-sects (e.g., Holldobler & Wilson, 1990), but aggregation is also observed inmany other invertebrate (e.g., Waters, 1959; Heip, 1975) and vertebrate (e.g.,Graves & Duvall, 1995; Domeier & Colin, 1997) taxa. In the Class Am-phibia, aggregation is well known in larvae (Brodie & Formanowicz, 1987;Glos et al., 2007; Smith & Awan, 2009) but is less recognized in adult an-imals. Amongst the poison dart frogs of the neotropics (family Dendrobati-dae), aggregations of adult frogs have very rarely been reported.

Animals aggregate for a variety of ecological and behavioural reasons.Aggregation can occur as a result of shared resource use (i.e., light, water,or food concentrated in one area), as a result of sociality, aggression, di-rect competition, or as a byproduct of reproductive strategies (Wells, 1977;Graves & Duvall, 1995; Karvonen et al., 2000; Knopp et al., 2008). In ad-dition, some animals aggregate to reduce parasite loads and/or risk of pre-dation through dilution amongst group members (Hamilton, 1971; Arnold& Wassersug, 1978; Graves et al., 1993; Glos et al., 2007; Smith & Awan,2009). In his classic paper ‘Geometry for the Selfish Herd’, W.D. Hamilton(1971) gives the hypothetical example of frogs jumping towards each otherto avoid predation, diluting their individual risk until they ‘quickly collect inheaps’ (p. 296).

Hamilton’s (1971) example was purely theoretical, but empirically, amonganurans, tadpoles will often aggregate in response to predation risk (Brodie& Formanowicz, 1987; Glos et al., 2007; Smith & Awan, 2009), or forthermal benefits (Espinoza & Quinteros, 2008). Newly metamorphosedAnaxyrus species have also been reported to aggregate to reduce desicca-tion amongst individuals (Heinen, 1993), as well as reduce risk of individualpredation (Arnold & Wassersug, 1978; Graves et al., 1993). Except for repro-ductive reasons (Lynch & Wilczynski, 2006; Swanson et al., 2007; Knopp etal., 2008; and reviewed in Wells, 1977), reports of aggregation behaviour inadult anurans are rare and, thus, to our knowledge, Hamilton’s (1971) classichypothetical example using adult frogs remains largely untested.

The Talamanca striped rocket frog (Allobates talamancae Cope) is a small(male SVL = 17–24 mm, female SVL = 16–24 mm), non-toxic member ofthe family Dendrobatidae which feeds on small arthropods in the leaf-litterof primary and secondary lowland wet forests in Central America (Savage,2002). Aggregation behaviour in this frog, like most other dendrobatids, has

Aggregation behaviour in Allobates talamancae 361

Figure 1. Photograph of four grouped captive A. talamancae sitting centimetres apart fromeach other in a terrarium.

not been described before. We observed in the laboratory that captive A. ta-lamancae individuals will tend to aggregate in multi-individual groups (Fig-ure 1), and are not evenly distributed in a terrarium. This observation led usto the purpose of this study, which was to determine if A. talamancae signifi-cantly aggregates in its natural habitat and to elucidate a possible explanationfor aggregation in this understudied species of dendrobatid frog.

2. Materials and methods

2.1. Study area

This study took place at the Institute for Tropical Ecology and Conserva-tion’s Bocas del Toro Biological Station, and in the tropical lowland wetforests in the surrounding area of Boca del Drago on Isla Colon in theBocas del Toro archipelago in western Panama (lat. 9◦20′40.7′′N, long.82◦30′54.1′′W).

2.2. Experiment 1: Dispersion of frogs along a 20 m transect in a lowlandforest habitat

In the morning of July 6, 2009, five 20-m transects were laid out in a lowlandforest habitat with fairly homogeneous leaf litter that had been cleared of themajority of under-story vegetation. Non-overlapping transects were laid outat randomly chosen angles of 110, 340, 224, 88 and 148◦ by placing a metre


tape on the ground. Each transect was then walked and the location of everyA. talamancae within a 1.0 m distance from either side of the transect wasnoted by placing a small piece of flagging tape at its location. Twenty randomnumbers between the values of 0.0 and 20.0 were then chosen along eachtransect using a random number table, and the distance from these points tothe location of each frog was measured, as prescribed by the point-quarterdistance method (Krebs, 1989). The point-quarter distance method was usedto determine the dispersion pattern of frogs along each transect. A ratio ofvariance/mean < 1 indicated a uniformly distributed population, a ratio equalto 1 indicated a randomly distributed population, and a variance/mean ratio >

1 indicated a clumped or aggregated population.

2.3. Experiment 2: Dispersion and nearest neighbour analysis of frogs ina 500 m2 plot in a lowland forest habitat

On July 8, 2009, a 50 × 10 m plot was laid out in the same forest as Ex-periment 1, by placing metre tape along the ground to mark out the plotboundaries. Two observers, evenly spaced along the 10 m, walked the lengthof the plot, and dropped a small flag at every location where a frog wasobserved. The distance from each flag to both its neighbour flag, and its sec-ond nearest neighbour was then measured to be used in a nearest neighbouranalysis of patterns of dispersion in this plot, where the mean distance tonearest neighbour and the mean distance to second nearest neighbour wascalculated and compared using a paired t-test (α = 0.05). In addition, acrossthis 50 × 10 m plot, 125 4.0 m2 quadrants were placed and the number offrogs found in each quadrant was counted to be used to calculate indices ofdispersion in the plot. The variance/mean ratio of number of frogs was cal-culated, as well as Morisita’s Index of Dispersion (Morisita, 1962; Graves etal., 1993), which was compared against an index value indicating random-ness, using a Chi-Squared test, with α = 0.05.

2.4. Experiment 3: Directional response of frogs to ‘predation risk’ in aterrarium habitat

Eight A. talamancae (N� = 2, N� = 6) were collected through manualsearch efforts, July 1–2, 2009, in a lowland tropical wet forest. These animalswere placed in a 113.6 l terrarium with approximately 2 cm of dried leaf littercollected from the same forest habitat from which the frogs were collected.


The leaf litter was homogenously distributed so as to provide a similar habitatthroughout the terrarium. A small plate with a quarter of a ripe banana wasplaced in the centre of the terrarium to attract frugiverous insects to providea food source for the anurans. A mesh screen was then placed over theterrarium to prevent the frogs from escaping, and the habitat was evenlymisted with water for moisture. Prior to experimentation, the animals weregiven a week to habituate in the terrarium.

For the experiment, a misted, empty 38 l terrarium was placed in a sepa-rate area that provided homogenous environmental characteristics (i.e., lightlevel) throughout the enclosure. Two randomly chosen frogs were placed onone side of the terrarium, while the other side of the terrarium was left bare.A third frog (focal frog) was then placed directly in the centre of the terrar-ium, facing neither the other two frogs nor the bare side of the terrarium.After approximately 15 s of acclimatization, an observer then brought hishand directly down on top of the middle frog as if to catch the animal. Thisaction was repeated up to five times until the focal frog initiated a flight re-sponse. The direction of flight, either towards the other two frogs or the bareside of the terrarium, was recorded. This procedure was repeated for eachof the eight captive frogs. In all cases, only the focal frog exhibited a flightresponse, with the other two frogs not moving during the simulated ‘attack’.To determine if there was an effect of what side of the terrarium the two frogswere placed, the whole experiment was repeated with the two frogs placedon the opposite side of the terrarium (rotate from ‘position 1’ to ‘position 2’).The number of frogs that fled to either direction was summed and comparedusing logistic regression (α = 0.05), with the binomial response variable be-ing presence or absence of focal frog on side of terrarium, after ‘predation’threat.

3. Results

3.1. Experiment 1: Dispersion of frogs along 20 m transects in a lowlandforest habitat

A total of eighteen A. talamancae (adult females, males, and subadults)were counted in this experiment. No frogs were located along transect 5.As such, this transect was excluded from any further analysis, and onlyresults for transects 1–4 are presented in Figure 2. The number of frogs


Figure 2. The locations of A. talamancae on 20 m line transects in a lowland forest in Bocasdel Toro, Panama (diamonds represent A. talamancae individuals). Variance: mean ratios arepresented for each transect, showing a significantly clumped population distribution in each

case (index of dispersion > 1.0).

found along each transect varied between 2–7 individuals. In each case, thevariance/mean ratio > 1.0, indicating their distributions were significantlyclumped, and that the frogs were aggregating in the wild (Figure 2).

3.2. Experiment 2: Dispersion and nearest neighbour analysis of frogs ina 500 m2 plot in a lowland forest habitat

The mean distance from one individual A. talamancae to another in the500 m2 plot was 1.02 ± 0.10 m, which was significantly less than the meandistance to its second nearest neighbour, 2.69 ± 0.14 m (paired t102 = 9.37,p < 0.0001), indicating that frogs were not uniformly distributed throughoutthe plot (Figure 3). When the number of frogs within quadrants was calcu-lated, frogs were found to be in a significantly clumped or aggregated distrib-ution throughout the plot (Figure 4), with a variance/mean ratio of 1.33, anda Morisita’s Index of Dispersion of 1.80, which was significantly differentfrom random (χ2

124 = 165.75, p < 0.01).

3.3. Experiment 3: Directional response of frogs to ‘predation risk’ in aterrarium habitat

When A. talamancae were ‘threatened’ by a simulated predation event, 7out of 8 frogs jumped to the side of the terrarium where two other frogswere located, versus the bare side of the terrarium (Figure 5). When the twofrogs were placed on the other side of the terrarium, and the experimentwas repeated, 6 out of 8 frogs jumped to the group of frogs present in the


Figure 3. Mean ± SE distances from each A. talamancae to its nearest neighbour and sec-ond nearest neighbour in a 500 m2 plot in lowland forest habitat in Bocas del Toro, Panama.Asterisks indicate significant differences between means (paired t102 = 9.37, p < 0.0001).

terrarium (Figure 5). In both cases, significantly more frogs chose to flee tothe side of the terrarium where conspecifics were present, versus the side ofthe terrarium without frogs (position 1: χ2

14 = 10.12, p < 0.001; position 2:χ2

14 = 4.19, p < 0.05).

4. Discussion

Allobates talamancae were observed to significantly aggregate in naturalhabitats. Our preliminary observations in captivity revealed that frogs werefound in mixed-sex groups of 2–5 individuals (Figure 1), and in our exper-iments in nature we found mostly groups of two or three individuals. Ourobservation of aggregation behaviour in this dendrobatid contrasts with re-sults by Summers (2000) that indicate encounters between A. talamancaeindividuals in the wild are relatively rare. Aggregation in post-metamorphicAnaxyrus cognatus in both captivity and the wild has been reported (Graveset al., 1993), but other than reports of breeding aggregations (Wells, 1977;Lynch & Wilczynski, 2006; Swanson et al., 2007) there are very few pub-lished cases of aggregation by adult anurans in either captivity or the wild.Tinbergen (1963) famously asked the question: How does the behaviour ofan animal promote its ability to survive and reproduce? Here, we discussand critique possible answers to this question of adaptivity in regards to ourobservations of aggregation behaviour in A. talamancae.


Figure 4. Map of 500 m2 plot in a lowland forest habitat in Bocas del Toro, Panama, withdots showing the location of each A. talamancae frog found (N = 53) (x- and y-axes arein m). The 4.0 m2 grid shown indicates the location of each of the 125 4.0 m2 quadrantsused in dispersion analyses (grid lines added after for visual effect). Frogs were significantlyaggregated in the plot (variance/mean ratio = 1.33; Morisita’s Index of Dispersion = 1.80,

χ2124 = 165.75, p < 0.01).


Figure 5. The directional movement of A. talamancae frogs (y-axis) when threatened bya ‘predator’ in experimental terrarium trials. Significantly more frogs moved towards con-specifics (‘frogs present’ vs. ‘frogs absent’) when threatened. (**position 1: χ2

14 = 10.12,p < 0.001; *position 2: χ2

14 = 4.19, p < 0.05). Position 2 differs from position 1 by thelocation of frogs in the terrarium moving 180◦.

Aggregation as an antipredator behaviour is likely adaptive, and it is welldocumented in a variety of taxa, ranging from insects, birds and fish to batsand ungulates (reviewed in Hamilton, 1971; Vulinec, 1990). According toHamilton (1971), animals will selfishly move closer to conspecifics whenthreatened by predation, so as to minimize their individual potential contactwith predators, which is more likely when one is alone or on the margins ofa group. Although Hamilton (1971) initially outlined his theory with adultfrogs, aggregation as an antipredator behaviour has primarily been empir-ically demonstrated only in larval anurans (Brodie & Formanowicz, 1987;Spieler & Linsenmair, 1999; Glos et al., 2007; Smith & Awan, 2009). Arnold& Wassersug (1978) do, however, hypothesize that aggregation behaviour inpost-metamorphic Anaxyrus boreas is an antipredator strategy, in which the‘selfish herd’ (Hamilton, 1971) effect applies. This was also found to be truewith post-metamorphic Anaxyrus cognatus, which aggregated in responseto conspecific chemical cues and appeared to flee predators faster when ingroups than alone (Graves et al., 1993). Our result that the vast majority(75–87.5%) of Allobates talamancae tested will significantly move to an ag-gregation of conspecifics when threatened by a ‘predator’ suggests a simi-lar antipredator function may exist in this leaf litter frog; thus, it providessome of the first preliminary empirical evidence of the ‘selfish herd’ effectthat Hamilton (1971) theoretically described in adult frogs. Aggregation mayseem to be a counter-intuitive antipredator strategy, as predators may be able


to see larger groups more easily. However, aggregation as an antipredatorresponse is selfishly selected for by the individual, not the group or species(Williams, 1966; Hamilton, 1971) and, thus, the individual benefits of riskdilution indicate that aggregation may still be selected for as an antipredatorresponse.

There is also a possibility that aggregation by frogs may not be a ‘self-ish herd’ effect (Hamilton, 1971), but rather function to confuse a predator(Miller, 1922), as has been shown for aggregations of marine insects (Tre-herne & Foster, 1981) and Daphnia (Milinski, 1979). Large groups of preyitems all moving together make it more difficult for a predator to chose whichprey item to attack (Miller, 1922), and as such, predator latency to attackis increased, and attacks are generally reduced in number (Milinski, 1979).The effectiveness of aggregations in producing a predator confusion effectis dependent on the size of the aggregation (Miller, 1922; Milinski, 1979),and without actual predation data, we do not know what that effective sizemight be for A. talamancae. Poulin et al. (2001) however found no Allobatesspecies present in an avian predator’s diet (that readily ate other neotropicalfrog and lizard species) and while this could be due to a number of reasons,it might be suggestive of an effective predator-mediated aggregation in A. ta-lamancae.

Regardless of the specific reason for antipredator aggregation (i.e., dilu-tion, Hamilton, 1971; or confusion, Miller, 1922), the development of thisbehaviour would likely be especially beneficial for A. talamancae, as unlikeits more colourful dendrobatid cousins, this species has no known protectiveskin toxins (Savage, 2002) and, thus, selection should favour development ofalternative predator avoidance and antipredator measures.

Aggressive and competitive behaviours, often tied with reproductivestrategies, can sometimes result in multiple individuals of a species cominginto contact with each other. This is well documented in a variety of verte-brate taxa, including mammals (Fryxell, 1987), birds (Karvonen et al., 2000),fish (Domeier & Colin, 1997) reptiles (Graves & Duvall, 1995) and amphib-ians (Wells, 1977). Aggressive behaviours between individual A. talamancaefrogs in our study were never observed in either captive or natural environ-ments, and this is in concurrence with the observation that neither male norfemale A. talamancae appear to show aggressive behaviour (such as fight-ing) to each other (unlike many closely related Dendrobates and Phyllobatesspecies) (Summers, 2000). A. talamancae males have been shown to not be


subject to mate guarding, nor is there evidence of female preference for a par-ticular male (Summers, 2000). This tolerance of multiple individuals aroundmates suggests a reproductive strategy in A. talamancae that could be con-ducive to aggregation, though Summers (2000) did not note the occurrence ofthis phenomenon in his study. Breeding aggregations and lekking behaviourare common in anurans that are explosive breeders (Wells, 1977; Swansonet al., 2007; Knopp et al., 2008; Smith & Awan, 2009), but have not beenreported to such a degree in prolonged breeders that reproduce throughoutthe year (Knopp et al., 2008). While A. talamancae is a prolonged breeder, itis possible that some modified form of lekking or reproductive aggregationcould still occur, and this hypothesis is consistent with our observations ofmultiple females and males in the same group in captivity. We were unable toreadily identify males and females in the field without catching them, and soinformation on the sexual make-up of individual groups of frogs found in theforest is not available. A possible inconsistency with reproductive behaviourbeing a driving force behind observed aggregations in our study is the factthat while being overall prolonged breeders, most related Allobates speciesreportedly mate more regularly in the rainy season, generally between No-vember and May (Juncá et al., 1994; Juncá, 1998; Lima et al., 2002). Thisstudy was conducted in early July, past that period of increased reproduc-tive activity, and we observed no mating occurring during the study (G.R.H.,personal observation).

Animals will often aggregate around a particular environmental resource,such as food or water. Classic examples of this phenomenon include theaggregation of birds and mammals around watering holes in east Africa(Ayeni, 1975), and the annual congregation of bear (Ursus arctos) predatorsto streams of spawning salmon and trout prey in North America (Mattson& Reinhart, 1995). Among amphibians, however, predator attraction to preyin frogs does not appear to be a significant determinant in predicting overallspatial patterns of predatory frogs in the wild (Hammond et al., 2007). Inaddition, reported aggregations of post-metamorphic Anaxyrus cognatus didnot occur as a response to environmental conditions, but rather as a result ofattractions to conspecifics (Graves et al., 1993). It is unlikely that attractionto resources was a significant influencing factor in the aggregation of A. tala-mancae in this study. Although we cannot rule out environmental differenceswithin microhabitats definitively, we made every effort to ensure a homoge-nous environment in the wild and aggregation towards any one part of thenatural habitat was not observed.


We have here presented evidence of aggregation behaviour occurring in awild adult neotropical dendrobatid leaf-litter frog. Quantitative observationsof this sort for the family Dendrobatidae are extremely rare, and our resultsthat aggregation may be an antipredator response for A. talamancae suggesta significant adaptive value for this under-reported behaviour in adult frogs.

Acknowledgements

We foremost thank Dr. E.D. Brodie, Jr. (Utah State University) for many insightful discus-sions and providing both constructive feedback on earlier versions of the manuscript andongoing support to G.R.H. We also thank the Utah State University Herpetology Group, par-ticularly B.G. Gall, for providing valuable comments on an earlier version of the manuscript.Dr. A. Beulig (New College of Florida, ITEC) and Dr. J. Loudon (University of Colorado–Boulder, ITEC) provided valuable advice and constructive feedback while in Panama, andA. Dixon helped with the field collection of frogs. We also especially acknowledge B. Bjeldefor assistance in data collection, care of captive frogs and providing useful feedback. Dr.B.H. Aukema (University of Minnesota) and Dr. S. Durham (Utah State University) providedvaluable statistical advice, and B.A. Rowland provided grammatical expertise. The commentsof the editor and an anonymous reviewer contributed greatly to the production of this man-uscript. This project was logistically and financially supported by the Institute for TropicalEcology and Conservation (ITEC).

References

Arnold, S.J. & Wassersug, R.J. (1978). Differential predation on metamorphic anurans bygarter snakes (Thamnophis): social behavior as a possible defense. — Ecology 59: 1014-1022.

Ayeni, J.S.O. (1975). Utilzation of waterholes in Tsavo National Park (East). — Afr. J. Ecol.13: 305-323.

Brodie Jr., E.D. & Formanowicz Jr., D.R. (1987). Antipredator mechanisms of larval anurans:protection of palatable individuals. — Herpetologica 43: 369-373.

Domeier, M.L. & Colin, P.L. (1997). Tropical reef fish spawning aggregations: defined andreviewed. — Bull. Mar. Sci. 60: 698-726.

Espinoza, R.E. & Quinteros, S. (2008). A hot knot of toads: aggregation provides thermalbenefits to metamorphic Andean toads. — J. Therm. Biol. 33: 67-75.

Fryxell, J.M. (1987). Lek breeding and territorial aggression in white-eared kob. — Ethology75: 211-220.

Fryxell, J.M. (1995). Aggregation and migration by grazing ungulates in relation to resourcesand predators. — In: Serengeti II: dynamics, management, and conservation of anecosytem (Sinclair, A.R.E. & Arcese, P., eds). University of Chicago Press, Chicago,IL, p. 257-273.

Glos, J., Erdmann, G., Dausmann, K.H. & Linsenmair, K.E. (2007). A comparative studyof predator-induced social aggregation of tadpoles in two anuran species from westernMadagascar. — Herpetol. J. 17: 261-268.


Graves, B.M. & Duvall, D. (1995). Aggregation of squamate reptiles associated with gesta-tion, oviposition, and parturition. — Herpetol. Monogr. 9: 102-119.

Graves, B.M., Summers, C.H. & Olmstead, K.L. (1993). Sensory mediation of aggregationamong postmetamorphic Bufo cognatus. — J. Herpetol. 27: 315-319.

Hamilton, W.D. (1971). Geometry for the selfish herd. — J. Theor. Biol. 31: 295-311.Hammond, J.I., Luttberg, B. & Sih, A. (2007). Predator and prey space use: dragonflies and

tadpoles in an interactive game. — Ecology 88: 1525-1535.Heinen, J.T. (1993). Aggregations of newly metamorphosed Bufo americanus — tests of two

hypotheses. — Can. J. Zool. 71: 334-338.Heip, C. (1975). On the significance of aggregation in some benthic marine invertebrates. —

In: Proceedings of the 9th European Marine Biology Symposium, Oban, 2–8 October1974, p. 527-538.

Heppner, F.H. (1974). Avian flight formations. — Bird Banding 45: 160-169.Holldobler, B. & Wilson, E.O. (1990). The ants. — Belknap, Cambridge.Juncá, F.A. (1998). Reproductive biology of Colostethus stepheni and Colostethus march-

esianus (Dendrobatidae), with the description of a new anuran mating behavior. —Herpetologica 54: 377-387.

Juncá, F.A., Altig, R. & Gascon, C. (1994). Breeding biology of Colostethus stepheni, a den-drobatid frog with a nontrasported nidicolous tadpole. — Copeia: 747-750.

Karvonen, E., Rintamäki, P.T. & Alatalo, R.V. (2000). Female–female aggression and femalemate choice on black grouse leks. — Anim. Behav. 59: 981-987.

Knopp, T., Heimoverta, M., Kokko, H. & Merila, J. (2008). Do male moor frogs (Ranaarvalis) lek with kin? — Mol. Ecol. 17: 2522-2530.

Krebs, C.J. (1989). Ecological methodology, 1st edn. — Harper & Collins, New York, NY.Lima, A.P., Caldwell, J.P. & Biavati, G.M. (2002). Territorial and reproductive behavior of

an Amazonian dendrobatid frog, Colostethus caeruleodactylus. — Copeia: 44-51.Lynch, K. & Wilczynski, W. (2006). Social regulation of plasma estradiol concentrations in

a female anuran. — Horm. Behav. 50: 101-106.Mattson, D.J. & Reinhart, D.P. (1995). Influence of cutthroat trout (Oncorhynchus clarki) on

behaviour and reproduction of Yellowstone grizzy bears (Ursus arctos), 1975–1989. —Can. J. Zool. 73: 2072-2079.

Milinski, M. (1979). Can an experienced predator overcome the confusion of swarming preymore easily? — Anim. Behav. 27: 1122-1126.

Miller, R.C. (1922). The significance of the gregarious habit. — Ecology 3: 122-126.Morisita, M. (1962). Id — index, a measure of dispersion of individuals. — Res. Popul. Ecol.

4: 1-7.Poulin, B., Lefebvre, G., Ibáñez, R., Jaramillo, C., Hernández, C. & Rand, A.S. (2001). Avian

predation upon lizards and frogs in a neotropical forest understorey. — J. Trop. Ecol.17: 21-40.

Savage, J.M. (2002). The amphibians and reptiles of Costa Rica: a herpetofauna between twocontinents, between two seas. — University of Chicago Press, Chicago, IL.

Smith, G.R. & Awan, A.R. (2009). The roles of predator identity and group size in the an-tipredator responses of American toad (Bufo americanus) and bullfrog (Rana catesbina)tadpoles. — Behaviour 146: 225-243.

Spieler, M. & Linsenmair, K.E. (1999). Aggregation behavior of Bufo maculates tadpoles asan antipredator mechanism. — Ethology 105: 665-686.


Summers, K. (2000). Mating and aggressive behaviour in dendrobatid frogs from CorcovadoNational Park, Costa Rica: a comparative study. — Behaviour 137: 7-24.

Swanson, E.M., Tekmen, S.M. & Bee, M.A. (2007). Do female frogs exploit inadvertentsocial information to locate breeding aggregations? — Can. J. Zool. 85: 921-932.

Tinbergen, N. (1963). On aims and methods of ethology. — Z. Tierpsychol. 30: 410-433.Treherne, J.E. & Foster, W.A. (1981). Group transmission of predator avoidance behaviour

in a marine insect: the Trafalgar effect. — Anim. Behav. 29: 911-917.Vulinec, K. (1990). Collective security: aggregation by insects as a defense. — In: Insect

defenses: adaptive mechanisms and strategies of prey and predators (Evans, D.L. &Schmidt, J.O., eds). State University of New York Press, Albany, NY, p. 251-288.

Waters, W.E. (1959). A quantitative measure of aggregation in insects. — J. Econ. Entomol.52: 1180-1184.

Wells, K.D. (1977). The social behaviour of anuran amphibians. — Anim. Behav. 25: 666-693.

Williams, G.C. (1966). Adaptation and natural selection, a critique of some current evolu-tionary thought. — Princeton University Press, Princeton, NJ.

aggregation behaviour in a neotropical dendrobatid frog ... · variety of taxa, amongst amphibians...

Documents